Return Home

Related Projects

Mugshot

The aim of this project is to demonstrate the capabilities of the SAFE AS face recognition system on low resolution images from the internet.  The SAFE AS team has developed a web interface and client-server platform to run face recognition on database of around 4000 labelled, low-resolution images obtained from trolling the internet and from the Labeled Faces in the Wild public dataset provided by University of Massachutsetts.  When a query image is given, the labelled dataset images are ranked in order of similarity and presented to the user.

Mobile Mugs

The aim of this project is to extend the application domain to mobile phones.  The advantages of the face recognition algorithm developed from the research at SAFE AS include its speed, robustness to environmental and pose factors and that it works well on low resolution images.  Given its low computational requirements it is well suited to mobile applications, particularly for photos taken in varying conditions and via low quality mobile phone cameras.  A prototype is being developed on the iPhone.

Intelligent CCTV for Proactive Security

Video surveillance systems have attracted worldwide attention since they were used to such great effect to track the movements of the four suicide bombers in the days before their attack on the London Underground in July 2005. Despite their usefulness, most current surveillance systems only provide reactive security by enabling the analysis of activities after the terrorist attack has already occurred — what is needed is proactive security to help prevent future attacks. Intelligent Closed-Circuit TV (ICCTV) systems use powerful computers to analyse the video feeds to assist human operators to detect events of interest as they occur — an example might be recognising the face of a suspected terrorist in a crowded railway station. In this project, we plan to run long term trials of advanced ICCTV technologies in important and sensitive public spaces such as major ports and railway stations, so that we can address operational and capability deficiencies in current ICCTV systems. The field trial component of the proposal in conjunction with established vendors will ensure that the ICCTV research is focussed on operational and real-world deployment issues in the Australian context. The research component funded by NICTA aims to improve the counter-terrorism capabilities and sensitivities of ICCTV systems whilst reducing the false alarm rate.

This project, "Intelligent CCTV for Proactive Security," was funded in 2006-2008 by the National Security Science and Technology Unit of the Department of Prime Minister and Cabinet.

Advanced Surveillance to Protect Critical Infrastructure

Intelligent Video Surveillance using CCTV is a necessary tool for agencies to
ensure public safety and protect critical infrastructure. Commercial ports,
Railway stations and other critical infrastructure around Australia are at
constant risk from security incidents that can put the public at risk and halt
operations. The key objective in this project is to apply, develop, and deploy
advanced technologies for real-time video analysis and presentation to actively
identify and track people and vehicles (including small boats). Furthermore,
3D video analysis from calibrated cameras will be used for incident detection
and anomalous behaviour identification.

This project, "Advanced Surveillance to Protect Critical Infrastructure," isfunded in 2009-2011 by the National Security Science and Technology Unit of the Department of Prime Minister and Cabinet.

Armadillo:  Open Source C++ Linear Algebra Library

Armadillo is a C++ linear algebra library (matrix maths) aiming towards a good balance between speed and ease of use. It's being developed as part of the SAFE AS project in order the provide a solid backbone for computationally intensive experimentation, while at the same time allowing for relatively painless transition of research code into production environments.

Exhibitions

• Demonstrated at IFSEC Security Trade Show (May 2009, Birmingham, UK)
  
• Demonstrated at IDEX Defence Trade Show (February 2009, Abu Dhabi, UAE)
  
• Demonstrated at ASIS Security Trade Show (September 2007, Las Vegas, USA)
  
• Demonstrated at CeBIT (May 2007, Sydney)

Invited Talks and Keynote Presentations 

• Keynote Address: Lovell, B.C., The Future of Security Surveillance:  Reliable Person and Vehicle Recognition, FutureCCTV09, Raffles Hotel, Singapore May 19-21, 2009
  
• Invited 1-Day Workshop: Lovell, B.C., Intelligent CCTV: Tricks of the Trade, FutureCCTV08, Singapore April 18, 2008
• Keynote Address: Lovell, B.C., Intelligent CCTV Project with Queensland Transport, FutureCCTV08, Raffles Hotel, Singapore April 16-17, 2008
• Keynote Address: Abbas Bigdeli, Intelligent CCTV Project with Queensland Transport, Transit Security and Infrastructure Design, Brisbane April 1, 2008

Awards and Recognition

• Brian Lovell elected President of the International Association for Pattern Recognition (IAPR), 2008
  
• Brian Lovell elected Fellow of the IAPR, 2008
  
• Most Cited Paper Award from Digital Signal Processing (Elsevier) in 2007 for Dr. Conrad Sanderson's paper: "Identity verification using speech and face information"